CFD Simulations Better understand, analyze, and optimize components and systems

At CC FNUM, different software packages are used to perform flow simulations. In addition to this, numerical methods are developed for increased simulation efficiency and to better fulfil specific simulation requirements. The results then represent the basis for optimizing flow conditions. 

Comparison of vortex strengths at large downstream distances between the commonly used SST model and the coupled Reynolds stress model.
Development of a coupled Reynolds stress model for the accurate simulation of vortex structures. Here: tip vortex of a NACA0012 wing.
By means of a novel approach to modeling multiphase flows, it is possible to determine cavitation regions computationally. The upper row of figures shows simulation results which indicate the cavitation region (yellow) for a pump at different operating points. These are in good agreement with the experimental observations in the lower row of figures.
Simulation of a pump impeller using the Lattice Boltzmann Method.
Tip gap vortex structures (colored with vorticity) and reduced total pressure contour plot for a centrifugal compressor. The results for two different RANS turbulence models are presented. Left picture: Full Reynolds stress model Right picture: Standard SST k-omega

Why CFD simulations?

Computational fluid dynamics (CFD) is ideally suited for the analysis and optimization of flows in individual components and entire systems. The effect of geometric modifications can be predicted quantitatively with high accuracy, even in cases where an absolute prediction is bound to higher uncertainties. Detailed parameter studies are easier to implement in numerical simulations than in experiments.

Complex multiphysics modelling is enjoying increasing popularity and is used for a wide variety of flows. Commercially available software is not always able to capture the phenomena of interest with sufficient accuracy. Therefore, specific CFD models and solvers are developed at CC FNUM to fill such gaps. Here are some examples: 

  • highly specialized turbulence models for the calculation of complex vortex structures 
  • novel multiphase flow approaches for the analysis of cavitation in pumps and turbines  
  • GPU-based Lattice-Boltzmann solvers for the detection of turbomachinery instabilities 
  • application of fluid-structure interactions in highly dynamic components and systems. 

With the continuing rapid growth in computing capacity, the future belongs to computational fluid dynamics. The CC Fluid Mechanics and Numerical Methods has been able to steadily increase the computing capacity of its own cluster over the last years. Currently, more than 200 processors with a total of more than 1800 cores and 7.6 TB RAM are used to perform increasingly complex flow simulations. This allows models with up to 100 million nodes to be processed. In order to be able to store the resulting data, the storage capacity has been expanded to over 130 TB.